Memory Controller Read Threshold Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory systems face challenges in maintaining quality of service (QoS) as they approach the end of their lifespan, particularly due to the distribution of fail-bits (FB) affecting error correction and read operations, leading to reduced performance and potential premature replacement.
Innovation Solution
A memory system and method that perform a test read operation using multiple read threshold entries to determine and select the best threshold for decoding data, partitioning the data set into good and bad sets, and optimizing the read level table (RLT) to improve QoS by adapting to NAND conditions and reducing fail-bit impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple read threshold entries are used for test read operations, then the quality of service (QoS) is improved through better data decoding, but the device complexity increases due to additional threshold management and partitioning operations
Solution Approach 1:
The patent segments the read data set into good data set and bad data set based on decoding success, and segments the read threshold entries into good read threshold entries and bad read threshold entries. This segmentation allows the system to manage complexity by handling different data portions with appropriate threshold strategies.
Solution Approach 2:
The patent performs preliminary classification of read threshold entries by testing them on a read data set before actual use. The controller determines good read threshold entries through test read operations and partitioning, preparing optimized thresholds in advance for subsequent read operations.
2Manufacturing precision
If the read data set is partitioned into good and bad data sets, then the manufacturing precision of data decoding is improved, but the loss of time increases due to additional partitioning operations
Solution Approach 1:
The patent applies partial action by performing partitioning operations only on portions of data that are necessary. The read data set is partitioned into good and bad data sets based on decoding requirements, applying detailed threshold optimization only where needed rather than to the entire data set.
Solution Approach 2:
The patent changes the parameter of read threshold values dynamically by selecting different thresholds from the read threshold table based on data characteristics. The controller adjusts threshold parameters adaptively for good and bad data sets, optimizing decoding precision without requiring complete reprocessing.
3Reliability
If bad data set is set as new read data set for reprocessing, then the reliability of data recovery is improved, but the duration of action increases due to iterative read operations
Solution Approach 1:
The patent implements feedback by using results from test read operations to determine good and bad read threshold entries. The system feeds back this information to select appropriate thresholds for subsequent read operations on the bad data set, improving recovery reliability through adaptive learning.
Solution Approach 2:
The patent applies dynamics by making the read threshold selection adaptive and changeable based on data conditions. The controller dynamically selects thresholds from the read threshold table based on the specific characteristics of the bad data set, rather than using fixed thresholds, enabling efficient recovery without excessive iteration.
Data Source
AI summary
A memory system includes a memory device and a controller. The controller performs a test read operation on a read data set of the memory device, using multiple read threshold entries and determines which are good read threshold entries based on results of the read operation. The controller selects a best read threshold entry among the multiple read threshold entries based on a result of the test read operation, partitions the read data set into a good data set decodable by the best read threshold entry and a bad data set undecodable by the best read threshold entry, and sets the bad data set as a new read data set.